Systems biology-based analysis of cell-free systems

Abstract

Cell-free expression systems are becoming increasingly widely used due to their diverse applications in biotechnology. Despite this rapid expansion in adoption, many aspects of cell-free systems remain surprisingly poorly understood. Systems biology approaches make it possible to characterize cell-free systems deeply and broadly to better understand their underlying complexity. Here, we review recent systems biology studies that have provided insight into cell-free systems. We focus on characterization of the cell-free proteome, including its dependence on preparation protocol and host strain, as well as the cell-free metabolome and the relationship of endogenous metabolism to system performance. We conclude by highlighting promising future research directions.

Graphical Abstract

Introduction

Over the course of evolution, cells have perfected the ability to process incoming information from their surrounding environment and modify their internal physiological state by altering intracellular levels of small molecules and proteins (their metabolome and proteome, respectively). Synthetic biology has harnessed and repurposed this ability for many biotechnological and bioindustrial applications, most often by using live cells as a host to execute complex functions. However, such “whole-cell” approaches can have profound limitations, from the metabolic burden of cell growth interfering with those functions to the limited robustness of cells in non-laboratory environments.

Cell-free expression (CFE) systems have emerged as a promising alternative to whole cells for a number of applications. CFE systems use a cell’s native transcription and translation machinery in an in vitro, open environment where even cytotoxic products can be expressed, overcoming numerous challenges inherent to using whole cells[1]. This machinery can be obtained via purification of recombinant elements (PURE)[2] but is more widely acquired via cellular extracts or lysates, as this approach is typically more economical. CFE systems have historically been used to address fundamental scientific questions; a renewed interest in their practical capabilities has led to their application for biosensors[3–5], therapeutics[6] and manufacturing novel proteins[1,7]. However, expanding the applications of CFE systems—especially on an industrial scale—requires significant advances such as boosting productivity, reproducibility, and robustness.

Escherichia coli-based extracts are the most widely used CFE systems, with well-established protocols and high productivity. Eukaryotic CFE systems prepared from yeast, wheat germ extract, rabbit reticulocyte lysate, and other sources have also been developed, though their protein productivity is significantly lower than that of E. coli-based extracts[8,9]. An overview of a typical workflow for production of E. coli extracts (as well as subsequent reactions) is presented in Figure 1, though we note there is often considerable variation in protocol details, including growth medium, inducers, lysis settings, etc. depending on the target application of the system[10–12]. Additionally, even under otherwise identical conditions and protocols, extracts prepared from different host strains or by different people can exhibit substantial variability[13].

Figure 1:

Overview of cell-free system protocols, from preparation to reaction. Preparation of lysate for cell-free expression typically entails harvesting cells at the mid-exponential growth phase, when the transcription and translation machinery is most active. After a series of wash steps, the cells are lysed via methods such as sonication or homogenization. The resulting extract may then be directly used for protein expression, though post-lysis processing of the extract via a “runoff” reaction and dialysis are often used to improve expression from endogenous promoters. In cell-free reactions, the bacterial lysate is supplemented with amino acids, nucleotides, salts, metabolites, and energy sources (typically phosphoenolpyruvate, or PEP). A DNA template (linear or plasmid) is then added to the assembled reaction reagents to initiate protein expression, which may be the end goal of the reaction or which may catalyze other chemical reactions. The protein yield of a reaction can be characterized by various methods, with one common approach being to measure the fluorescence of a reporter like green fluorescent protein (GFP) with a plate reader.

As a result, cellular extracts are often considered a “black box” consisting of a poorly understood and characterized mix of native machinery[14]. The scarcity of investigation into the underlying details of the “black box” has hindered solving several issues associated with CFE systems, including batch-to-batch variability and limited reaction life. Numerous hypotheses have been formulated to explain these observations, including degradation of polypeptides produced during the reaction, depletion of metabolites, accumulation of toxic molecules, and changes in lysate pH[15–17]. Efficient tools to test these hypotheses could ultimately guide improvements to CFE systems.

Accordingly, systems biology—a “top-down” scientific approach where measurements and models are made at the level of the whole biological system—can be a powerful tool to characterize and understand CFE systems. The lysate and the entire CFE system can be investigated at every level of cellular or molecular regulation using advanced experimental techniques. Here, we review recent investigations of CFE systems using systems biology, mainly within the past five years. We focus first on proteomics studies, then summarize metabolomics studies, and finally discuss promising future research directions.

Section 1: Proteomics-based investigations of CFE systems

Proteomics—the measurement of the complement of proteins in a biological system—is a powerful tool for biological investigation. E. coli cells have been characterized extensively using proteomics, including quantitative measurements of proteins under varying environmental conditions[18,19]. Lysate-based CFE systems also have a proteome: the lysate preparation is designed to preserve the proteins of the transcriptional and translational machinery, leading to non-specific preservation of many other proteins in the lysate. Thus, unlike PURE systems—which are composed of purified proteins and ribosomes—crude extract-based CFE systems do not have known concentrations of a known set of proteins[20].

Accordingly, proteomics has been used to study CFE systems, often with the intent to support practical goals like identification of pathway bottlenecks during protein expression[21], analysis of post-translational modifications[20], or improvement of the yield of biosynthetic products[22,23]. Early studies of proteins in CFE systems used 2-D gel electrophoresis to identify changes in the protein composition over the course of a reaction[24] and for lysates with different translational activities[25]. Later proteomic analysis of E. coli-based lysates determined that only around a quarter of the total potential E. coli proteome is present in lysates derived from cells extracted at mid-exponential growth phase[21], with missing proteins including membrane transporters that are removed along with the cell membrane and proteins expressed only during other growth phases. With continuing technological advances in proteomics, it has now become possible to measure not just the identities of the proteome constituents, but also their absolute concentrations[26].

The host strain of the lysate has been shown to have a significant impact on its proteome. For example, studies of commercial systems like myTXTL and standard extract preparations like S30 have demonstrated that lysates from different host strains have different profiles of metabolic pathway enzymes. Major catabolic pathways are overrepresented in the S30 lysate (prepared from E. coli A19)[21], whereas myTXTL (prepared from E. coli BL21 Rosetta2) shows a shortage of a glucose transporter enzyme[27] that renders glucose an inefficient carbon source.

Cell growth conditions can also affect the proteome and final protein productivity of CFE systems, with important parameters including harvest time[28], mode of cultivation (e.g., bioreactors or shake flasks), and lysate preparation methods[29]. The proteome of E. coli changes rapidly during growth[19], thus harvesting cells at different stages of growth leads to extracts of differing protein composition[28]. Interestingly, lysate harvested at stationary phase can show protein productivity comparable to that of prototypical lysates harvested from mid-exponential phase.

Beyond basic studies of CFE systems, proteomics has also been used to characterize and optimize CFE for specific applications. For example, while CFE has been used to manufacture a variety of proteins such as cytokines and urokinases[30], it has been challenging to synthesize more complex proteins[20]. Proteomic analysis has provided important information regarding these limitations, including the identification of proteases at high levels in lysate as a potential target for removal to enhance protein production[21]. Also, elongation has been identified as the rate-limiting step of translation in CFE, with supplementation of purified elongation factors helping to enhance protein synthesis[28].

Proteomics analysis of CFE systems has also facilitated improved production of small molecules such as pyruvate[22] and phenol[23]. In cell-free metabolic engineering studies, proteomics data has been used to guide rerouting of metabolic pathways to optimize production of the target molecules. Related to these efforts, in vivo studies indicating a causal link between proteome allocation efficiency and the metabolic phenotype of E. coli cells[31] suggest that it is also important to study CFE systems at a metabolic level—i.e., via metabolomics. Such efforts could deepen our understanding of how these systems work and expand the potential targets for optimization to improve system function[32].

Section 2: Metabolomics-based investigation of CFE systems

Metabolomics studies of CFE systems have great potential to complement the insights resulting from proteomics analysis. While proteomics can identify which proteins are present in the lysate and their relative levels, it does not measure their activity or any downstream indicator of that activity. In particular, many of the proteins expected to maintain function and utility in both in vivo and cell-free environments are metabolic enzymes that act on small molecules in the CFE system, making metabolomics approaches particularly relevant.

Metabolomics and targeted metabolite profiling studies have already provided significant insight into the biochemical reactions happening during CFE. An early metabolism-focused study of CFE systems used mass spectrometry for real-time monitoring of metabolites involved in the production of dihydroxyacetonephosphate (DHAP) via glycolysis[33]. This approach identified the impacts of cofactors and purified enzymes added during the reaction and guided improvements to the system. More recently, mass spectrometry was used to quantify glycolytic compounds[34] in the cell extract and to demonstrate that a new growth medium improves energy metabolism and redox balance[35]. Nuclear magnetic resonance (NMR) spectroscopy has been used to identify energy bottlenecks[36]—notably, ATP consumption by ribosomes—and small-molecule (e.g., ethanol) indicators of lysate activity[37].

Other recent efforts have helped to broaden our understanding of the metabolic implications and impacts of CFE. Differences in protein production between lysates made with different lysis post-processing steps or from cells grown in different media were found to have corresponding differences in metabolite levels[35,38]. When monitoring metabolite profiles during protein expression, it was found that while metabolite levels change throughout the reaction, changes are most significant early on and proceed even after protein synthesis has stopped[39]. Importantly, these collective efforts have helped highlight the importance of the “endogenous metabolism” in CFE systems resulting from the metabolic enzymes extracted with the expression machinery; this endogenous metabolism seems to have a major impact on protein production, and its impacts on metabolite profiles are much greater than the impacts of protein expression itself.

The results of these studies also guided efforts to supplement CFE reactions to enhance protein production. Addition of small molecules such as putrescine and beta-alanine to the reaction improved protein production nearly three-fold[38], helping support the causal link between metabolite levels and lysate productivity. However, it is worth noting that the impacts of metabolite supplementation can be hard to predict, even when driven by metabolomics data, as they are highly dependent on the concentrations of the metabolites being supplemented and the effects of different supplements are not additive. Enzyme supplementation was also explored, though of the enzymes tested, few significantly enhanced protein production and their resulting impacts on metabolite profiles were only slight[39].

Taken together, metabolomics studies to date have provided useful insight into the metabolic underpinnings of CFE systems and even allowed for improvement in their protein production. Identification of the fact that the dominant metabolic changes in CFE systems arise from the lysate’s endogenous metabolism rather than protein production could help guide future efforts to engineer and optimize CFE. However, more work remains, as these efforts have not yet yielded a sufficiently detailed understanding of the system to enable truly rational engineering of the system.

Section 3: Future directions

Unraveling the complexities of CFE systems through omics-based systems-level analysis is a relatively novel research niche with great potential for further exploration. Although individual metabolomics and proteomics analyses have provided useful insights into E. coli-based lysates, they have yet to be integrated into a single, multi-omics study better capable of capturing the complexity of the system[40]. Other omics techniques, such as lipidomics, could also be used to supplement these characterizations.

Another important avenue of research would be to characterize protein levels over time (as a reaction proceeds or after lysate preincubation), as previously done at the metabolite level[39]. Identifying the core proteome of the lysate specifies which biological pathways are available but cannot predict how usage of those pathways will change over time or whether proteins involved in those pathways get degraded. Metabolomic analysis has shown that endogenous lysate activity causes most metabolic changes in the extract; a similar effect may exist for the proteome. To this end, the ability to characterize proteome activity over time would be quite useful but currently faces technological limitations.

A noteworthy limitation of proteomic analysis is its inability to determine the activity of the measured proteins. For example, different degrees of cell lysis can result in different total protein content in the lysate. While higher lysate protein content is generally expected to be preferable[11,41], lower protein content (resulting from gentler lysis) sometimes enables higher protein synthesis than a more harshly lysed extract with higher content[39], perhaps due to higher sonication energy inputs leading to protein degradation or denaturation. This demonstrates the importance of distinguishing between protein concentration and activity, as well as the potentially significant value of high-throughput methods to assay for the activity of individual proteome constituents.

Targeted alterations to the lysate preparation protocol—for example, using anaerobic fermentation during lysate preparation to alter cellular metabolism[42]—have succeeded in improving lysate productivity but the underlying causes, potentially related to changes in the metabolome and proteome, have not been fully explored[42]. An omics-based study of those new lysates could reveal specific changes in the metabolome and proteome that may have contributed to the increase in yield[35].

Finally, omics analysis of CFE has thus far been restricted to lysates derived from a very small number of E. coli strains, but it could be particularly beneficial for the study of lysates derived from other strains and other organisms. A multi-omics analysis could help explain, for example, why certain strains, such as DH10B, are not ideal for lysate production. More importantly, it could help improve the productivity of lysates derived from more complex organisms, such as yeast, that have yet to match the productivity of E. coli-based lysates. This would advance the development of CFE systems that enable high-yield production of complex proteins that, for instance, may undergo post-translational modifications absent in simpler organisms.

These and other research directions are promising options to apply systems biology to expand our understanding of cell-free systems and enable its broader and more effective use in biotechnological and bioindustrial contexts.

Highlights.

• Cell-free systems are a promising technique in biotechnology and bioindustrial applications to overcome limitations inherent to working with whole cells.

• Proteomics has provided useful insight into the protein-level variability in cell-free systems and its relationship to system performance.

• Metabolomics has revealed the importance of the endogenous metabolism in cell-free lysate and its impact on system performance.

• Further applications of these and other systems biology approaches are promising avenues to enable future improvements in cell-free systems and their applications.